Model Overview
Alibaba · Active Model
- Long-Context
- API Available
- Production Ready
The Qwen3.5 native vision-language Flash models are built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. Compared to the...
- Developer
- Alibaba
- Release Date
- February 25, 2026
- Context Window
- 1,000,000 tokens≈ 1,333 words
- API Access
- Publicly AvailableIntegrate via official API
- Input Cost
- $0.07per million tokens
- Output Cost
- $0.26per million tokens